Adobe’s Agentic AI Emerges as Organizing Layer Across Productivity, Creativity, Customer Experience

Agentic AI Is Becoming the Organizing Layer of Adobe’s Strategy

Bottom line

Agentic AI appears to be a company-wide strategic transition for Adobe, not merely a feature enhancement or a new product category. Management is framing the shift similarly to Adobe’s earlier transition from on-premise software to SaaS: every major software company will need to evolve toward agentic software, and Adobe intends to lead that transition across creativity, productivity, and customer experience. ¹

The likely long-term outcome is an agentic Adobe platform in which users interact through either traditional interfaces or conversational agents, while Adobe’s applications, models, enterprise data, and workflows operate together in the background. In customer experience, the more ambitious end state is an end-to-end system of record and orchestration layer that can identify prospects, generate and optimize content, execute campaigns, and continuously improve customer engagement. ¹²

Management has not provided a standalone Agentic revenue forecast, so the financial magnitude cannot yet be quantified precisely. However, its strategic importance is evident from the extent to which Adobe is embedding it across products, go-to-market operations, customer deployments, and its stated long-term growth strategy. ³⁴

How large a factor will Agentic be?

1. It is intended to span the entire company

Adobe is explicitly positioning Agentic across its three principal businesses:

• Productivity: Acrobat and Reader are being developed into AI-powered document and knowledge platforms, with conversational interfaces, document analysis, knowledge bases, interactive reports, and audio summaries. ³⁵
• Creativity: Firefly, Creative Cloud, Photoshop, Premiere, and other applications are incorporating AI agents that can handle complex or repetitive creative workflows. ⁶
• Customer experience: Adobe CX Enterprise is being positioned as an agentic system for customer-experience orchestration across the customer lifecycle. ²

This breadth matters because Agentic is not confined to one experimental application. Adobe is applying the same strategic concept across its largest product categories, while using its existing distribution, data, workflows, and customer relationships as the foundation. ⁷

2. It is central to Adobe’s next growth cycle

Adobe’s leadership has connected Agentic directly to the company’s next decade of growth, stating that AI will expand Adobe’s reach, customer value, and impact. ⁸ The incoming CEO also described his mandate as driving Adobe’s growth and leadership across creativity, productivity, and customer experience “in the era of Agentic software.” ³

The strategy is therefore broader than monetizing AI credits. Adobe is attempting to use agents to:

• increase product engagement;
• expand the addressable user base;
• make complex capabilities accessible to nonexperts;
• deepen enterprise adoption;
• increase the value of existing workflows; and
• extend Adobe’s products into external interfaces and ecosystems. ⁷³⁴

Adobe already has substantial distribution on which to build: more than 1 billion monthly active users across its businesses, more than 20,000 global enterprise customers, and customer-experience platforms that deliver more than 1 trillion experiences annually. ⁸⁹ These assets give Agentic a potentially much larger economic base than a standalone AI application would have.

3. Early operating indicators show meaningful adoption, although not yet a separately disclosed revenue stream

The disclosed indicators suggest that customers are beginning to use Agentic capabilities at scale:

• Creative freemium monthly active users exceeded 100 million and grew by more than 70% year over year. ¹⁰
• Firefly ending ARR across the Firefly application and credit packs increased 40% sequentially. ¹⁰
• Ending ARR grew more than 20% year over year for Adobe Experience Manager and Agentic web apps, Adobe GenStudio, and Adobe Experience Platform and applications. ¹⁰
• Adobe doubled the number of paid customers for its brand-visibility solutions sequentially. ¹⁰
• CX Enterprise CoWorker became generally available in June, and Adobe reported 1,700 customers and early adopters for the offering. ¹¹
• Acrobat AI Assistant monthly active users doubled quarter over quarter. ⁵

These figures are encouraging, but they should not be treated as proof that Agentic has already become a major independent revenue contributor. Adobe reports many of these products within broader subscription categories, and its FY2026 guidance provides total revenue and subscription targets rather than a distinct Agentic revenue line. ¹⁰⁴ The more defensible conclusion is that Agentic is already influencing adoption, usage, product packaging, and enterprise sales, while its ultimate contribution remains difficult to isolate.

The long-term path

Phase 1: Embed agents into existing Adobe applications

The first stage is to make Adobe’s installed products more intelligent without requiring customers to abandon familiar workflows.

Adobe is adding conversational and agentic interfaces to Reader, Acrobat, Express, Creative Cloud, Photoshop, and Premiere. ³⁶ In Acrobat, the productivity agent can turn complex documents into interactive reports, summary slides, and audio formats, while knowledge-base and analyzer features can extract insights across large document collections. ⁵

In Creative Cloud, the Creative Agent is designed to orchestrate complex and repetitive workflows, and Adobe has expanded it into flagship applications including Photoshop and Premiere. ⁶ This approach should allow Adobe to preserve the value of its existing applications while increasing the amount of work performed inside them.

Phase 2: Become interface- and model-agnostic

Adobe’s intended agentic architecture is not limited to Adobe’s own user interface or a single proprietary model.

Management identified three customer requirements:

1. Interface flexibility: customers should be able to use conversational or traditional interfaces, including external interfaces such as ChatGPT, Claude, or Copilot, while still accessing Adobe functionality. ¹
2. Model flexibility: Adobe should match the best available model to the particular task rather than force customers to use one model. ¹
3. Enterprise context: applications should use the customer’s data and context to improve accuracy, predictability, and usefulness while reducing hallucination risk. ¹

Adobe is already extending Acrobat capabilities to platforms including ChatGPT, Chrome, Microsoft Edge, and WhatsApp. ³ Strategically, this suggests that Adobe is seeking to become the trusted execution and workflow layer behind multiple interfaces, rather than insisting that every interaction begin inside an Adobe application.

That strategy could broaden Adobe’s reach, but it also creates a risk: if external platforms own the customer interaction and Adobe supplies only underlying functionality, Adobe may capture less of the economics and customer relationship. The filing excerpts establish Adobe’s interface strategy, but they do not yet establish how revenue sharing, customer ownership, or distribution economics will work. ¹³

Phase 3: Connect agents to enterprise data, workflows, and governance

For enterprise customers, the end product is not simply a chatbot. Adobe is building an agentic operating layer that combines:

• proprietary customer and brand intelligence;
• enterprise data and context;
• model orchestration;
• governed workflow execution;
• open ecosystem connectivity; and
• enterprise security and testing. ⁹

Adobe CX Enterprise illustrates this direction. It combines Adobe Experience Platform and customer-engagement applications, GenStudio for the content supply chain, and Experience Manager and agentic web applications for brand visibility. ⁹

The CX Enterprise CoWorker is designed to synthesize information from Adobe and third-party applications while executing complex marketing and customer-engagement workflows. ¹¹ That is a more valuable proposition than content generation alone because the agent is intended to move from recommendation to coordinated action.

Adobe is also building a brand-visibility offering that combines Semrush’s AI-visibility intelligence with Adobe’s content-optimization capabilities, using a database of nearly 300 million real-world AI search prompts to assess how brands appear across platforms such as ChatGPT, Google AI Mode, Microsoft Copilot, and Perplexity. ⁹ This points toward an environment in which Adobe helps enterprises not only create content, but also measure and optimize how that content performs across generative search systems.

Phase 4: Reach an end state of autonomous, closed-loop customer and creative workflows

The most ambitious end state is an end-to-end agentic system that can understand business objectives, access relevant data, create or modify content, execute governed workflows, measure outcomes, and optimize continuously. Adobe says it has built a practical path from current applications to such an end-to-end agentic AI system. ⁹

In customer experience, the intended destination is especially clear: Adobe CX Enterprise is described as an agentic system of record for orchestrating the entire customer lifecycle, from identifying and engaging prospects through conversion and lasting loyalty. ²

In creativity, the parallel end state is an agent that can coordinate multi-step production across formats and applications while preserving professional control, brand consistency, and creative intent. Firefly Enterprise already includes custom brand-safe models, reusable node-based workflows, mass-scale production, and campaign-performance simulation. ¹²

In productivity, the end state is an AI-powered knowledge and document platform that can understand large collections of enterprise information and convert them into usable analysis, decisions, and finished work. ⁵

Thus, the likely end state is not “Adobe adds an AI assistant to each application.” It is Adobe applications becoming coordinated agents that perform work across documents, creative assets, marketing systems, enterprise data, and external platforms. This is an analytical interpretation of management’s stated direction, rather than a separately quantified company target. ¹⁹⁷

Why Adobe may be well positioned

Adobe’s competitive advantage in this transition is based on accumulated assets rather than AI models alone.

Management points to:

• decades of expertise in PDF and document workflows;
• deep domain knowledge across imaging, design, video, photography, illustration, animation, and 3D;
• more than 1 billion monthly active users;
• more than 400 billion PDFs opened with Acrobat annually; and
• Adobe’s position as a provider of customer-experience orchestration for more than 1 trillion experiences per year. ⁷³⁶⁸

Those assets can help Adobe provide the context, workflow integration, trust, and distribution that generic models lack. Adobe is also emphasizing commercially safe AI, enterprise governance, security, compliance, and predictable outputs, which are particularly important for large organizations. ⁶⁵

The strategic value of Agentic may therefore come less from owning the best general-purpose model and more from owning the domain-specific workflows, enterprise context, and trusted execution environment around those models. Adobe’s stated model-flexibility strategy reinforces this interpretation. ¹

Key financial and strategic questions

Revenue expansion versus cannibalization

Agentic could expand Adobe’s opportunity by attracting new users, increasing usage, and enabling more enterprise workflows. Adobe reported that its creative freemium user base exceeded 100 million and that business-professional and consumer monthly active users exceeded 900 million, suggesting substantial distribution for AI-enabled upselling. ⁵²

However, Agentic could also alter how software is purchased and consumed. If customers increasingly interact through external agents, they may use fewer traditional application interfaces or expect more usage-based pricing. Adobe has not disclosed how its pricing, packaging, or revenue recognition will evolve across these scenarios. The excerpts therefore support a strong strategic commitment but not a precise forecast of incremental margins or revenue. ¹¹⁰⁴

Cost and margin discipline

Agentic workloads can require substantial model-inference, storage, data-processing, and customer-support resources. Adobe says it is focused on delivering AI securely, accurately, and cost efficiently, and FY2026 targets assume a non-GAAP operating margin of approximately 45%. ⁷⁴

That suggests management is attempting to scale Agentic while preserving Adobe’s software economics. Nevertheless, the provided disclosures do not quantify AI-related infrastructure costs, gross margins by AI product, or the degree to which higher usage will be offset by pricing and automation. Margin durability remains an important unresolved question.

Execution complexity

Adobe must coordinate multiple models, interfaces, applications, customer data sources, permissions, and enterprise workflows. Management is investing in forward-deployed engineering and value-realization teams to help customers implement and capture value from Agentic software. ¹³¹¹

This is strategically sensible because enterprise adoption often depends on implementation and measurable business outcomes, not just product availability. It also means that part of Adobe’s Agentic growth may require services-like deployment effort, which could increase sales complexity and near-term operating intensity even if it strengthens long-term customer retention. This latter implication is an analytical risk, not a disclosed financial forecast. ¹³¹¹

Competitive pressure

Adobe’s strategy assumes that customers will value Adobe’s domain expertise, integration, security, and workflow orchestration even when powerful general-purpose models are available through competing interfaces. Management believes Adobe can lead in agentic creativity, productivity, and customer experience, but the excerpts do not provide evidence that this competitive outcome is assured. ¹³¹

The most important competitive question is whether Adobe remains the primary platform where work is executed, or whether external AI interfaces become the primary customer relationship while Adobe functions mainly as a specialized service provider. Adobe’s decision to support external interfaces addresses reach, but it may also increase dependence on those platforms. ¹³

Overall assessment

Agentic is likely to be one of the dominant strategic factors in Adobe’s next several years. It is being applied across every major product category, linked explicitly to the company’s next decade of growth, incorporated into enterprise sales and customer-success motions, and used to redefine Adobe’s products as coordinated systems rather than isolated applications. ³⁸⁴

The long-term path appears to be:

1. embed agents into existing Adobe products;
2. make those agents accessible through Adobe and third-party interfaces;
3. route tasks across the best available models;
4. ground outputs in enterprise data, brand context, and domain-specific workflows;
5. execute governed multi-step actions across Adobe and third-party systems; and
6. create closed-loop systems that generate, distribute, measure, and optimize content and customer experiences. ¹⁹⁷

The ultimate opportunity is substantial because Adobe already controls widely used applications, large user populations, valuable content and document workflows, and deep enterprise relationships. ⁹⁷⁸ But the financial outcome is not yet established: Adobe has shown promising adoption and usage indicators, while its disclosures still do not separate Agentic revenue, profitability, customer lifetime value, or infrastructure costs from the broader businesses. ⁵¹⁰⁴

The most credible conclusion is that Agentic will become the operating architecture of Adobe’s strategy, while its economic importance will be determined by whether Adobe can convert higher AI engagement into durable pricing power, broader enterprise workflows, and stronger customer retention without surrendering the interface or economics to external AI platforms.